OCt.,
1919
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
939
ORIGINAL PAPERS MODIFICATIONS OF PEARCE’S METHOD FOR ARSENIC By JOHN WADDELL
Received March 27, 1919
Owing t o the difficulty my students experienced in using the Pearce method for determining arsenic in a cobalt ore, I decided t o carry out some experiments. Pearce’s original method consisted in fusing t h e ore with sodium carbonate and potassium nitrate, extracting t h e fused mass with water, boiling the filtrate with excess of nitric acid t o drive off carbon dioxide, neutralizing with ammonia, precipitating the arsenate as silver arsenate, dissolving this precipitate in nitric acid, and titrating t h e silver with potassium or ammonium thiocyanate. Since silver arsenate is soluble not only in nitric acid and in ammonia, but also t o a slight extent in ammonium nitrate, the results are not quite satisfactory. Canby’ recommends neutralization of the nitric acid with zinc oxide in excess. He directs that should the zinc oxide cause precipitation of silica or alumina (Canby fused in a porcelain crucible) t o filter off, and t o the filtrate add more zinc oxide. Bennett2 criticizes Canby’s method, stating that i t is practically impossible t o get the zinc oxide to neutralize the nitric acid and that, therefore, silver arsenate is not completely precipitated. He himself recommends t h a t after fusion and filtration from the residue obtained by leaching in hot water the filtrate shall be acidified with acetic acid and boiled. After cooling, phenolphthalein is added as indicator and caustic soda solution t o slight alkalinity and then just enough very dilute acetic acid, drop by drop, until the color is discharged. Silver nitrate is then added a n d t h e arsenate titrated as usual. It may be noted t h a t Bennett read his burette only t o t h e first place of decimals and t h a t the quantity of potassium thiocyanate was small, ranging for the same quantity of arsenic between 8.3 cc. and 8.7 cc., the greater number of the determinations being 8.6 cc. This last number he probably considered nearest correct, for he took for each determination what was supposed t o be 0 . 0 2 j g. of impure arsenic trisulfide, and 8.6 cc. gave o . o z 4 7 z j g. of arsenic trisulfide, and 8 . 7 cc. gave 0.025013 g. It is t o be noticed t h a t Bennett, while he reads the burette t o the first place of decimals only, calculates t h e weight t o t h e fifth significant figure. Since silver acetate is only sparingly soluble, especially in a solution containing, as Rennett’s did, considerable sodium acetate, I think it probable t h a t not only 8.7 cc. but also 8.6 cc. was too high and that the 5.3 cc. may have been nearer right. Bennett gives no method for checking his results. A number of my students got entirely discordant results in using t h e two methods, and in fact got discordant results in determinations by the same method. In some cases, if not in most, in t h e Canby method, 1 2
Z. anal. Chem., 29, 187, abstracted in J . Chem. Soc., 08 (1890),920. J . A m . Chenz. Soc., 21 (1899),431.
the zinc oxide had been filtered off before precipitating with silver nitrate and as zinc nitrate gives a precipitate with sodium arsenate it was natural t o conclude t h a t arsenic had been lost in this way. Concordant results, however, were not obtained even when the zinc oxide was not filtered off and it seemed likely that t h e zinc arsenate was not all changed into silver arsenate. Even if t h e silver nitrate was added first and the zinc oxide afterwards, the results were still unsatisfactory. It seemed evident then t h a t the zinc nitrate reacted on silver arsenate and gave an error, and it was found by a qualitative test t h a t silver arsenate when digested with a concentrated solution of zinc nitrate became lighter in color, white zinc arsenate doubtless replacing silver arsenate. It seemed then reasonable t o conclude t h a t when in the Canby method any considerable excess of nitric acid was present, t h e zinc nitrate produced prevented the precipitation of all the arsenic as silver arsenate, and t h a t if t h e excess of nitric acid was very small -the results might be accurate. Accordingly, a solution of sodium arsenate, NasHAsOd.IzHzO, of approxima.tely 4 g. per 1. was tested. Four lots of I O O cc. each were boiled in Erlenmeyer flasks with j cc. of concentrated nitric acid. I n Flask A, zinc oxide was added and allowed t o stand for about an hour before addition of silver nitrate solution. The precipitate of silver arsenate did not come down at once, but after a few minutes a considerable amount was precipitated. I n Flask B, t h e silver nitrate solution was added first and afterwards zinc oxide. No precipitate appeared for an hour or two. In Flask C, the nitric acid, after cooling, was neutralized and the solution made slightly alkaline, using phenolphthalein as indicator, by dissolving pure caustic soda directly from the stick, so that carbon dioxide would not be taken up from the air. Then t h e color of the phenolphthalein was discharged by adding very dilute nitric acid drop by drop; excess of silver nitrate was added and a small quantity of zinc oxide. I n Flask D, the solution was made alkaline as before and dilute acetic acid was added until there was a drop or two in excess, after which the precipitation was made with silver nitrate. It may be mentioned t h a t t h e silver arsenate precipitate in A and B was dark brown; and often, though not always, t h e precipitate from the nitrate solution was slightly darker than t h a t from the acetate solution. This did not seem t o be due either t o the contrast of color with t h e white zinc oxide on the one hand, or t o the presence of white silver acetate with silver arsenate on the other. So far as I could detect, t h e composition did not vary with slight differences of shade. The four flasks were allowed t o stand for a fortnight, being occasionally shaken so t h a t the zinc oxide should have every opportunity t o react on t h e nitric acid. The filtration was done in Gooch crucibles with asbestos, and t h e crucibles after filtration were placed
940
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
in beakers; the Erlenmeyer flasks were washed out with about 2 0 rc. hot nitric acid ( I : I) which was then brought into the crucibles in the beakers. The silver arsenate dissolved easily, the solution was made up t o ahout 2 0 0 cc., the crucible was removed, the asbestos being left in the beaker, and after addition of 5 cc. saturated solution of ferric alum, the silver was titrated with ammonium thiocyanate. The results were as follows: AMMONIUM THIOCYANATE REQUIRED cc. A 18.90 B 20.40 1 3 9 . IO C. ,9 0 7-"F D 38.69
........................... ........................... ........... -.. ............................
It was thus seen t h a t with zinc oxide used t o neutralize a large amount of nitric acid the results were valueless. This was not, however, as Bennett suggests, because zinc oxide does not neutralize the acid, €or the filtrate from the mixture of zinc oxide and silver arsenate was neutral t o methyl orange, though a drop or two of very dilute nitric acid gave the red color. To test the method, I dissolved 1.5 g. of Kahlbaum's pure arsenious anhydride in nitric acid made up t o 500 cc. and took lots of 2 5 cc. ea"c. 'After I haddone some titrating myself I asked one of my students, Miss Lucille Corbett, t o do some determinations, which she carried out both on my solution and on one which she prepared for herself. In my determination by the Bennett method I added sodium acetate in considerable excess t o use up the nitric acid, while Miss Corbett added-caustic sods t o slight alkalinity and then acetic acid until slightly acid. Her figure thus obtained was less than mine and more nearly those of the Canby method which in this and subsequent tests was carried out in the modified style with only a slight excess of nitric acid t o be neutralized by zinc oxide. Miss Corbett and I used the same solution of thiocyanate and the same burette and took, respectively, 2 7 . 1 0 cc. and 2 7 . 2 0 cc. thiocyanate for 0.5 g. of silver nitrate. We used the second figure which gave I cc. thiocynate = o.002704 g. arsenic (Ag = 108; As = 75). One of Miss Corbett's determinations in each case was made with my solution and the other with her own. Another student, P. A. Poynton, using a different burette and a different thiocyanate solution which was, however, of almost exactly the same concentration as mine, made this determination also. Waddell.,
Canby Method cc. 20.98 21.00 21.05
.................
...................
Corbett Poynton..
.................
20.81
Bennett Method cc.
21.40 20.93 20.90 21 .OS
With my solution of thiocyanate, 21.00 cc. corresponded t o 75.68 per cent of arsenic, while with Poyn ton's 21.08 cc. gave the same percentage, the calculated value being 75.75. The next experiments were made with arsenious anhydride fused with ten or twelve times the amount of sodium peroxide, I n all the fusions which follow
Vol.
11,
No.
IO
sodium peroxide was used instead of a mixture of sodium carbonate and potassium nitrate, a nickel or iron crucible being employed. The fusion, after extraction with water, was made up t o a liter or half a liter as the case might be, and a sufficient quantity taken t o require a t least 14 cc. of the thiocyanate. Miss Corbett evidently lost some arsenic by volatilization but the results by the two. methods agreed closely. The freeing of the residue from arsenic, by extraction of the fused mass with water, is difficult. I n my experience, i t was not possible t o wash out all the arsenic when filtering through asbestos in a Gooch crucible. I filtered by preference through a large filter paper and washed till the washings were no longer alkaline. The residue should always be dissolved in hydrochloric acid and tested with a current of sulfureted hydrogen. It is needless t o say t h a t delicate tests such as the Gutzeit in even its crudest form would be certain t o show arsenic in the residue, and are not worth trying. Canby Method Per cent As
.................. {:::E ...................76.15 70.91 ....................
Waddell. Corbett., Poynton
Bennett Method Per cent As 75.96 76.49 71.06 78.46
The average of my lowest Bennett and highest Canby is 75.64 per cent As, the calculated being 75.75 per cent. It may be noted t h a t in taking only 1 5 cc. of thiocyanate a difference of unity in the second place of decimals in reading the burette makes a difference of five in the second place of decimals in the percentage. It will thus be evident what degree of accuracy Bennett attained when his determination required only 8.6 cc. and was read merely t o the first place of decimals. Analyses were made of mispickel. Miss Corbett, Poynton, and I each fused separate lots of 0 . 2 g. and determined the arsenic in the whole amount. Afterwards the other two fused 0.5 g. of mispickel, made up to a definite volume, and took sufficient of the solution for the determination. The second line for each in the table gives these results. Canby Method Per cent As 22.91 23.00 (23.38 24.04 POyYItOn.. . . . . . . . . . . . 23.68
......... .. Corbett. ................
Waddell.
(
Bennett Method Per cent As 23.38 23.79 24.23 23 61
It will be noted t h a t one of Miss Corbett's results by the Canby method was the same as mine by the Bennett method. Three other students also made analyses of mispickel taken from the same bottle as before but a t a different time, their sample being sifted through a n 80-mesh sieve. They all used the same burette and the same thiocyanate solution, both b q e t t e and solution being different from those already mentioned. H. K.' Rowley had already in another laboratory done a number of analyses using Bennett's original method in which there was a large excess of sodium acetate, assuming t h a t his results there obtained were correet. He was asked t o compare his former procedure with t h e proposed modification; and t h e results given below indicate t h a t the suspicion t h a t Bennett's results
Oct., 1919
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
were too high was correct. R . J. Young and J. S. Stauffer used only the modified method. Canby Method Per cent As Rowley ..................... 24.92 Stauffer.. . . . . . . . . . . . . . . . . . . .24.39 Young ...................... 24.55
Bennett Method Per cent As
{ ;:: ;; 24.90 25.02
With the single exception mentioned, none of the students had attempted the analysis before and in the case of all it was part of their prescribed course. None of them was able t o spend enough time to carry out a complete research; but t o all, the difficulties were explained and the object t o be attained was pointed out and they carried out their analyses with the aim and in the spirit of research. Two other students, H. C. Boehmer and I. L. Sills, did some pioneer work over a year ago, chiefly in varying tests of Canby’s method, and though their results are not included in this paper, their work was as important as t h a t of the others. I may add t h a t in work like this, I think it is possible for students in the second year of their analytical course t o get some insight into the methods of research which will help them towards the initiative and judgment so much desired by the industries employing university students and graduates. SUMMARY
I-Bennett’s modification of Pearce’s method for arsenic, if carried out as he describes, is likely t o give too high results, unless arsenic has been lost by volatilization or otherwise, while Canby’s modification, if carried out as he describes, may be so low as t o be valueless. 11-Both methods may be modified t o give practically concordant results; and if duplicates, determined one by the Bennett method and the other by the Canby method, each modified as described, agree, then the result may be considered correct. 111-In the modification of Bennett’s method, any large amount of alkali is acidified with nitric acid, made slightly alkaline with pure caustic soda and very slightly acid with acetic acid, before precipitation of the arsenate as silver arsenate. In the modification of the Canby method the process is similar, except t h a t instead of acetic acid, nitric acid is added in very slight excess and, after addition of silver nitrate, this small excess -is neutralized by zinc oxide. IV-Various determinations are given illustrating the degree of accuracy. CHEMISTRY DEPARTMENT QTJSSN’S UNIVERSITY KINGSTON, ONTARIO
AN IMPROVED METHOD FOR DETERMINATION OF CARBON BY WET COMBUSTION, USING BARIUM HYDROXIDE AS ABSORBENT By P. 3,. HIBBARD Received March 10, 1919
For the determination of carbon by wet combustion, the writer has for more than a year used a modification and combination of previously published
941
m e t h o d s 1 ~ 2 ~ 3with > 4 ~ 6much success. The method is simple, convenient, inexpensive, rapid, and accurate. It is here presented, with the hope t h a t it may be of use t o others. Complete combustion is secured by suitable proportion and quantity of reagents. Carrying over of volatile acid fumes t o the absorbent is avoided by an efficient purifying train. Convenience and accuracy in use of barium hydroxide as absorbent for carbon dioxide are secured by a simple and efficient apparatus without difficult manipulation. The operation is briefly as follows: The substance is heated in a Kjeldahl flask with chromic anhydride and sulfuric acid whereby carbon is oxidized t o carbon dioxide which is carried into a solution of barium hydroxide by a current of purified air. After the reaction is completed the excess of barium hydroxide is determined by titration with standard hydrochloric acid. The amount of barium hydroxide neutralized by carbon dioxide measures the amount of carbon in the substance taken. DESCRIPTION OF APPARATUS
By reference t o Fig. I the various parts and arrangement of the apparatus may readily be discovered. First is a bubble tube A, containing a few drops of colored liquid as indicator of t h e speed of t h e air current. B is a large test tube filled with soda lime for purifying the incoming air. C is the regulating stopcock; D is a funnel tube with a long stem extending down t o the bulb of the Kjeldahl flask F. The end of this long stem is somewhat drawn out t o a small opening. Upon D, connected by a two-hole rubber stopper rests E, a graduated dropping funnel for measuring the reagents. F is a long neck, 300 cc. Kjeldahl flask, used for the combustion chamber. From this the gas passes through a glass tube t o the bottom of G, a large test tube drawn out a t the lower end and fitted with a pinchcock and rubber tube. This acts as the condenser t o remove most of the water from the gas. After each combustion is finished the water is drained out of this tube b y opening the pinchcock. From G the gas passes through a long tube down t o H, a 5 0 cc., wide-mouth flask containing about I O cc. of strong sulfuric acid. Upon H, which is fitted with a two-hole rubber stopper, rests a glass tube, I, filled with glass beads wet with sulfuric acid, t h e purpose of which is t o dry the gas. Upon I, connected by a rubber stopper, rests J , a similar tube filled with granulated amalgamated zinc, the function of which is t o remove sulfuric acid or other acid fumes from the gas. From the purifying tube J the gas passes through a long tube down t o the 500 cc. Florence flask K, which is connected by a two-hole rubber stopper t o the Meyer bulb tube L. The lower end of this bulb tube is bent so as t o almost touch the bottom of the flask. At t h e upper end the bend next t o t h e large bulb is partly straightened out so t h a t the bulb stands upright. Ames and Gaither, THISJOURNAL, 8 (1916), 1126. a Truog, I b i d . , 1 (1915), 1045. Schollenberger, I b i d . , 4 (1912), 436. 4 Gortner, Soil Science, 2 (1916), 395. 6 Brady, THIS JOURNAL, 6 (1914), 843. 1
*
